2▲ 2 ▼ 0
Some musings on phonotactics in Tirina
This public article was written by alynnidalar, and last updated on 16 Mar 2015, 15:10.
2. To AddAs you know, Bob, syllables can consist of up to three parts, the onset (first bit), nucleus (middle bit), and coda (last bit). In Tirina, as in many languages, the onset and coda are technically optional.
Onsets, which again are optional, can consist of either any single consonant, or one of a small number of consonant clusters, of which Tirina has very few.
Acceptable single consonants: /f l k m p d t r n j h s w ʔ/
Acceptable consonant clusters: /mr pr fr/
Single consonants are significantly more common, but /mr pr/ are still used with reasonable frequency. /fr/ is fairly rare, however, and is actually pronounced more like /ɸr/. You'll notice that all of the clusters take the form bilabial + /r/.
As a side note, remember that "t" and "d" are not actually [t] and [d], they're [t̪] and [d̪]. Also, "r" represents both [r] and [ɾ] (although when in a cluster, it's [r]).
The nucleus, which is not optional, consists of any single vowel. No polyphthongs exist in Tirina, so you're stuck with /a e i o u ɪ/.
Like the onset, a coda is optional. Also like the onset, it can consist of either a single consonant or a consonant cluster, but the options are much more limited here.
Acceptable single consonants: /l d t r n s ʔ/
Acceptable consonant cluster: /rn/
Codas are less common than onsets. It's reasonably rare to find a syllable that consists only of a nucleus and a coda; if there's a coda, there's frequently an onset as well. All of the codas are dental/alveolar.
"Same Consonant" Rule
Note: the following section is somewhat out of date.
Pretty simple rule: you can't have the same phoneme twice in the same syllable. So *lol isn't acceptable, *non sounds silly, and *prarn is just ridiculous. (And yet "contain" is dud. Rules can be broken, kids, they just aren't usually.)
Similarly, double letters are entirely disallowed. (and this time, I actually do mean entirely.) A syllable cannot end with the same phoneme that the next syllable begins with. (*anna, *pii, etc.) If, through some rule of derivation or conjugation or whatever, this occurs, one of two things happens:
- If the doubled letters are consonants, the second is replaced by a glottal stop.
- If the doubled letters are vowels, a glottal stop is inserted between them.
Thus, a word such as nadana (she is throwing) becomes ton'adana (she is not throwing) when negated, rather than *tonnadana.
However, it's perfectly acceptable to have the onset of a codaless syllable be the same as the onset of the next syllable, such as adader (feminine), which consists of the syllables /a/ /da/ /der/.
Random Other Stuff
A few other notes on what is/isn't allowed...
/ɪ/ (represented orthographically as the dotless i "ı") strongly prefers to be followed by a consonant, whether it be a coda or the onset of the next syllable. If it falls at the end of a word, it is always followed by the glottal stop.
/dt/ generally resolves to /d/.
Two liquids (r and l) are also quite rare to be found together. (/rl/, /lr/, or even /lVr/ and /rVl/ in some contexts) In most cases it'll resolve to one or the other.
I'm not entirely certain if stress really counts as part of phonotactics, but it's relevant and not likely to show up anywhere else, so why not? Generally speaking, stress falls in two places for stems:
- For words of one or two syllables: first syllable
- For words of three or more syllables: second syllable
Therefore, edir is pronounced /'e.dir/, mudolan is pronounced /mu.'do.lan/, and ayirdoya is pronounced /a.'jir.do.ja/. (man, to know, and horse, respectively)
However, there are exceptions. There's a number of reasons why these exceptions could occur--various derivational processes, compounding, loanwords, or simply trying to avoid stressing ı--so there's no easy way to point to a particular word and know that it has irregular stress, aside from simply memorizing them. When romanized, these irregular stresses are occasionally denoted by the use of an accent mark over the appropriate vowel. However, this is not always the case, and certain very common words are almost never written with these accent marks.
An example is fotikan, library. Ordinarily, one might assume that it would be pronounced */fo-'ti-kan/. However, it is actually /'fo-ti-kan/, and is sometimes written as fótikan to indicate this. Another example is the name of the dalar species. Although the stress irregularly falls on the second syllable, it is almost never written as dalár; it is such a common word that it is simply assumed everyone knows how to pronounce it.
One helpful rule of thumb is that stressing ı (which, again, represents [ɪ]) is avoided whenever possible. However, there's no way to predict whether the stress will fall before or after the ı, in multisyllable words. Again, these ones just have to be memorized.
- unvoiced stops and nasals followed by /r/ (i.e., /tr/, /dr/, and /nr/) generally end up assimilated to a bilabial (so /tr/ and /dr/ -> /pr/ and /nr/ -> /mr/).